An accurate methodology to identify the level of aggregation in solution by PGSE NMR measurements: The case of half-sandwich diamino ruthenium(II) salts

Article abstract of DOI:10.1021/om050145k

The utility of PGSE NMR measurements in determining hydrodynamic radii (rH) and volumes (V_H) of small- and medium-size molecules (3 A < r_H < 6 A) was evaluated by performing measurements for a variety of pure deuterated solvents and their solutions containing the internal standard TMSS [tetrakis(trimethylsilyl)silane] also in the presence of a variable concentration of 3BPh₄. It was found that accurate r _H and V_H values can be obtained by introducing in the Stokes-Einstein equation (D_t = kT/cπηr_H) not only the correct values for temperature (T) and viscosity (η) but, particularly, that for the c factor. PGSE NMR measurements were then applied to an investigation of the aggregation tendency of complexes [Ru(η⁶- cymene)(R₁R₂NCH₂CH₂NR ₁R₂)Cl]X (R₁ = R₂ = H, 1; R ₁ = H, R₂ = H, 2; R₁ = R₂ = Me, 3; X^- = PF₆^- or BPh₁¹) in both protic and aprotic solvents with a relative permittivity (ε_r) ranging from 4.81 (chloroform-d) to 32.66 (methanol-d ₄). Compounds 1 and 2 exhibited a remarkable tendency to aggregate through intercationic N-H...Cl and cation/ anion N-H...FPF ₅^- hydrogen bonds. In addition to ion pairs, ion triples and ion quadruples were also observed in solution. Compound 3, having no N-H moiety, showed less tendency to aggregate than 1 and 2, even though it also afforded ion quadruples in apolar and aprotic solvents. Relative anion-cation orientations and arene conformations were investigated by means of ¹H-NOESY and ¹⁹F, ¹H-HOESY NMR spectroscopy. The relative anion-cation position was well-defined, especially for compounds bearing the PF₆^- counterion, and was modulated by the nature of the N,N ligand. A progressive slackening of the contact aggregates was observed in the series 1-3 that led to a higher mobility of the anion, as indicated by the observation of less specific interionic NOEs.

Full text of DOI:10.1021/om050145k

3476
Organometallics 2005, 24, 3476-3486
An Accurate Methodology to Identify the Level of
Aggregation in Solution by PGSE NMR Measurements:
The Case of Half-Sandwich Diamino Ruthenium(II) Salts
Daniele Zuccaccia and Alceo Macchioni*
Dipartimento di Chimica, Universita` di Perugia, Via Elce di Sotto, 8-06123 Perugia, Italy
Received March 1, 2005
The utility of PGSE NMR measurements in determining hydrodynamic radii (rH) and
volumes (VH) of small- and medium-size molecules (3 Å < rH < 6 Å) was evaluated by
performing measurements for a variety of pure deuterated solvents and their solutions
containing the internal standard TMSS [tetrakis(trimethylsilyl)silane] also in the presence
of a variable concentration of 3BPh4. It was found that accurate rH and VH values can be
obtained by introducing in the Stokes-Einstein equation (Dt ) kT/cπηrH) not only the correct
values for temperature (T) and viscosity (η) but, particularly, that for the c factor. PGSE
NMR measurements were then applied to an investigation of the aggregation tendency of
complexes [Ru(η6-cymene)(R1R2NCH2CH2NR1R2)Cl]X (R1 ) R2 ) H, 1; R1 ) H, R2 ) H, 2; R1
-
-
) R2 ) Me, 3; X- ) PF6 or BPh4 ) in both protic and aprotic solvents with a relative
permittivity (ꢀr) ranging from 4.81 (chloroform-d) to 32.66 (methanol-d4). Compounds 1 and
2 exhibited a remarkable tendency to aggregate through intercationic N-H‚‚‚Cl and cation/
anion N-H‚‚‚FPF5- hydrogen bonds. In addition to ion pairs, ion triples and ion quadruples
were also observed in solution. Compound 3, having no N-H moiety, showed less tendency
to aggregate than 1 and 2, even though it also afforded ion quadruples in apolar and aprotic
solvents. Relative anion-cation orientations and arene conformations were investigated by
1
means of H-NOESY and 19F,1H-HOESY NMR spectroscopy. The relative anion-cation
-
position was well-defined, especially for compounds bearing the PF6 counterion, and was
modulated by the nature of the N,N ligand. A progressive slackening of the contact aggregates
was observed in the series 1-3 that led to a higher mobility of the anion, as indicated by
the observation of less specific interionic NOEs.
Introduction
5%, accurate values of rH can be derived once the three
other parameters of eq 1, T, c, and η, are correctly
evaluated. While T and η can be measured, the selection
of the proper c factor is critical. This is particularly true
for medium- and small-size molecules, for which c differs
significantly from both 4 (slip boundary conditions) or
6 (stick boundary conditions). To avoid the measure-
ment of T and η and the evaluation of c, an internal
standard of known van der Waals radius (rVdW) can be
used to determine the proportional constant between
Dt and rH of eq 1.4,5 Two assumptions are implicit in
this procedure: (a) rVdW is a good representation of rH
for the internal standard and (b) the proportional
constant for the standard and that for the sample are
the same and, since they are in the same solution and
experience the same T and η, this means that they have
the same c factor. Assumption (a) is usually considered
correct.6,7 Since the c factor substantially depends on
the size of the diffusing species, assumption (b) is only
respected if the standard and the sample have the same
or, at least, comparable dimensions. Otherwise, c has
to be evaluated. An acceptable semiempirical estimation
Diffusion NMR spectroscopy1 has proven to be a
powerful tool for determining molecular size in solu-
tion.2,3 The hydrodynamic radius (rH) of the diffusing
species can be estimated from the experimentally
determined self-diffusion translational coefficient (Dt)
by taking advantage of the Stokes-Einstein equation:
kT
cπηrH
Dt )
(1)
where k is the Boltzmann constant, T is the tempera-
ture, c is a numerical factor, and η is the solution
viscosity. Because the error on Dt is usually lower than
* To whom correspondence should be addressed. E-mail: alceo@
unipg.it. Fax: +39 075 5855598. Phone: +39 075 5855579.
(1) Hahn, E. L. Phys. Rev. 1950, 80, 580-594. Stejskal, E. O.;
Tanner, J. E. J. Chem Phys. 1965, 42, 288. Stilbs, P. Prog. Nucl. Magn.
Reson. Spectrosc. 1987, 19, 1. Price, W. S. Concepts Magn. Res. 1997,
9, 299. Price, W. S. Concepts Magn. Res. 1998, 10, 197. Johnson, C.
S., Jr. Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 203.
(2) For reviews on applications to organometallic chemistry, see:
Valentini, M.; Ru¨egger, H.; Pregosin, P. S. Helv. Chim. Acta 2001, 84,
2833. Binotti, B.; Macchioni, A.; Zuccaccia, C.; Zuccaccia, D. Comments
Inorg. Chem. 2002, 23, 417. Pregosin, P. S.; Martinez-Viviente, E.;
Kumar, P. G. A. Dalton Trans. 2003, 4007.
(4) Mo, H.; Pochapsky, T. C. J. Phys. Chem. B 1997, 101, 4485.
(5) Zuccaccia, C.; Bellachioma, G.; Cardaci, G.; Macchioni, A. Or-
ganometallics 2000, 19, 4663.
(3) For a recent review on the application in supramolecular
chemistry, see: Cohen, Y.; Avram, L.; Frish, L. Angew. Chem., Int.
Ed. 2005, 44, 520.
(6) (a) Bondi, A. J. Phys. Chem. 1964, 68, 441. (b) Rowland R. S.;
Taylor, R. J. Phys. Chem. 1996, 100, 7384.
(7) Edward, J. T. J. Chem. Educ. 1970, 47, 261.
10.1021/om050145k CCC: $30.25 © 2005 American Chemical Society
Publication on Web 06/04/2005

Half-Sandwich Diamino Ru(II) Salts
Organometallics, Vol. 24, No. 14, 2005 3477
tion level in solution has to be determined through
diffusion measurements, the use of an internal standard
becomes problematic since the average dimensions of
the species present in solution can easily change when
solvent or concentration vary and, consequently, the
standard cannot be suitable for all circumstances. In
such cases eqs 2 and 3 have to be used for evaluating c
and rH, respectively.
In this paper, we first show that treating Dt data
coming from PGSE NMR measurements with eq 3 leads
to a more accurate estimation of rH with respect to those
obtainable using eq 1 with c ) 4 or 6. Second, we report
the PGSE results, treated with the methodology based
on eq 3, pertinent to the aggregation of [(arene)Ru(N,N)-
Cl]X compounds (where N,N ) diamine ligands) in
various solvents. Finally, for the latter, relative anion-
cation orientations and the conformations of cymene,
determined by NOE NMR measurements, are also
illustrated.
Half-sandwich diamino ruthenium(II) salts were se-
lected in order to contrast their tendency to aggregate
with that of analogous diimine complexes, which is
remarkable.12 In addition, even though they have been
known for several decades,15 there has been a renewed
interest in them since Sadler and co-workers16 demon-
strated their anticancer activity. This activity seems to
be connected to the possibility of such compounds to
strongly and selectively bind N7 of guanine and its
derivatives, in a physiological environment. In-depth
studies16 have shown that in such binding a crucial role
is played by the covalent coordination of N7 on the metal
center, π-stacking between the nucleobase and the arene
of the complexes, and stereospecific H-bonding between
the hydrogen-bond donors (NH functionalities of Ru
complexes) and hydrogen-bond acceptors (carbonyl func-
tionalities on nitrogenous base). It appears reasonable
that H-bonding and π-stacking, which are responsible
for the favorable interaction between DNA and [(arene)-
Ru(N,N)Cl]X salts, could also lead to self-aggregation
of the latter in solution with the possible formation of
ion pairs or even higher aggregates.
Figure 1. Dependence of the c factor of a diffusing particle
on the hydrodynamic radius (rH) in methylene chloride-d2
according to eq 2.
of c can be obtained through eq 28 derived from the
microfriction theory proposed by Wirtz and co-workers,9
in which c is expressed as a function of the solute-to-
solvent ratio of radii.
6
c )
(2)
2.234
rsolv
rH
1 + 0.695
(
)
[
]
Combining eqs 1 and 2, the Stokes-Einstein equa-
tion, eq 3, is obtained in which Dt depends on both rH
and the hydrodynamic radius of solvent (rsolv).
2.234
rsolv
kT 1 + 0.695
(
)
[
]
rH
Dt )
(3)
6πηrH
An example of the dependence of c on rH is reported
in Figure 1 for methylene chloride-d2 (rsolv ) 2.49 Å).
Mononuclear organometallic compounds have a hydro-
dynamic radius that usually falls in the 3-6 Å range.
Therefore, according to Figure 1, c varies from 4.1 to
5.5. Consequently, if c ) 6 is introduced into eq 3, the
error on the derived rH spans from 9% to 32%.
In our laboratory, we have been interested in the
aggregation of transition-metal organometallics in solu-
tion for several years.10-13 We are trying to correlate
the effect of aggregation on the chemical reactivity14
with the intermolecular structure. For intermolecular
structure, we mean both the relative position of the
interacting moieties, determined by NOE (nuclear Over-
hauser effect) NMR experiments, and the level of
aggregation, estimated by PGSE (pulsed field gradient
spin-echo) NMR technique. Clearly, when the aggrega-
Results
Synthesis. Complexes [Ru(η6-cymene)(R1R2NCH2-
CH2NR1R2)Cl]X shown in Scheme 1 were synthesized
by the reaction of [Ru2(η6-cymene)2Cl2(µ-Cl)2] with the
appropriate ligand in methanol at room temperature by
adding a large excess of NaBPh4 or NH4PF6, which
caused the precipitation of the desired ruthenium salt.15
Due to their high solubility in methanol, complexes 2PF6
and 3PF6 were preferentially synthesized by 2BPh4 and
3BPh4 through anion metathesis using an equimolar
quantity of TlPF6 in methylene chloride.
(8) Chen, H.-C.; Chen, S.-H. J. Phys. Chem. 1984, 88, 5118.
Espinosa, P. J.; de la Torre, J. G. J. Phys. Chem. 1987, 91, 3612.
(9) Gierer, A.; Wirtz, K. Z. Naturforsch. A 1953, 8, 522. Spernol, A.;
Wirtz, K. Z. Naturforsch. A 1953, 8, 532.
(10) Macchioni, A. Eur. J. Inorg. Chem. 2003, 195. Macchioni, A.
In Perspectives in Organometallic Chemistry; Screttas, C. G., Steele,
B. R., Eds.; The Royal Society of Chemistry: Cambridge, 2003; pp 196-
207.
(11) Zuccaccia, C.; Stahl, N. G.; Macchioni, A.; Chen, M.-C.; Roberts,
J. A.; Marks, T. J. J. Am. Chem. Soc. 2004, 126, 1448.
(12) Zuccaccia, D.; Sabatini, S.; Bellachioma, G.; Cardaci, G.; Clot,
E.; Macchioni, A.; Inorg. Chem. 2003, 42, 5465.
(13) Zuccaccia, D.; Clot, E.; Macchioni, A. New J. Chem. 2005, 29,
430.
Of the three possible isomers of complexes 2X (Scheme
2), differing in the relative orientation of the R substit-
(15) Crabtree, R. H.; Pearman, A. J. J. Organomet. Chem. 1977, 141,
325.
(16) (a) Novakova, O.; Chen, H.; Vrana, O.; Rodger, A.; Sadler, P.
J.; Brabek, V. Biochemistry 2003, 42, 11544. (b) Chen, H.; Morris, R.
E.; Sadler, P. J. J. Am. Chem Soc. 2003, 125, 173. (c) Aird, R. E.;
Cumming, J.; Ritchie, A. A.; Muir, M.; Morris, R. E.; Chen, H.; Sadler,
P. J.; Jodrell, D. L. Br. J. Cancer 2002, 86, 1652. (d) Chen, H.;
Parkinson, J.; Morris, R. E.; Parson, S.; Coxall, R. A.; Gould, R. O.;
Sadler, P. J. J. Am. Chem Soc. 2002, 124, 3064. (e) Morris, R. E.; Aird,
R. E.; Murdoch, Piedad del Socorro; Chen, H.; Cumming, J.; Hughes,
N.; Parson, S.; Parkin, A.; Boyd, G.; Jodrell, D. L.; Sadler, P. J. J. Med.
Chem. 2001, 44, 1652.
(14) For a recent review on ion pairing effect in transition-metal
organometallic chemistry, see: Macchioni, A. Chem. Rev. 2005, in
press.

3478 Organometallics, Vol. 24, No. 14, 2005
Zuccaccia and Macchioni
Scheme 1
Table 1. Diffusion Coefficients (109Dt m2 s-1), van
der Waals Radii (rvdW), Experimental
Scheme 2
Hydrodynamic Radii (rH), and Experimental
conv
Hydrodynamic Radii at Convergence (rH
)
Obtained for Pure Deuterated Solvent (Er at 25 °C)
conv
Dt
rvdW
rH
(∆r/r %) rH
1
2
3
4
5
6
7
8
benzene-d6 (2.27)
chloroform-d (4.81a)
CD2Cl2 (8.93)
2.02 2.7
2.42 2.6
2.68
2.65
-0.7
1.9
3.37 2.49 2.46
-1.2
2-propanol-d8 (19.92) 0.47 2.40 2.86
20
2.0
22
25
3.4
3.08
acetone-d6 (20.56)
ethanol-d6 (24.55)
methanol-d4 (32.66)
DMSO-d6 (46.45)
4.15 2.49 2.54
0.94 2.25 2.75
2.12 1.99 2.48
0.61 2.63 2.72
3.10
2.76
uents, only the anti isomer was observed in solution.
The metal center in the latter is chiral and, conse-
quently, two enantiomers are present in solution in
equal amount.
Evaluation of the Accuracy of PGSE NMR Mea-
surements. (a) Determination of the Hydro-
dynamic Volumes of Solvents. The goodness of eq 3
was checked by measuring Dt for samples of pure
deuterated solvents following the decrease of the signal
due to the small percentage of nondeuterated solvent
as a function of the gradient strength (G) increases
(Experimental Section).
The viscosity values introduced in the modified
Stokes-Einstein equation, eq 3, were known from the
literature;17 they were corrected to take into account
both the actual temperature17 and the deuteration of
the solvents.18 The Dt of water, which is necessary to
calibrate the gradient intensities, or better, to measure
the proportionality constant to obtain Dt (Experimental
Section), was also interpolated at the corrected temper-
ature.19 The Dt values agree with those reported in the
literature.18 Experimental rH and van der Waals radii
(rvdW) of the investigated solvents are reported in Table
1. For aprotic solvents (Table 1, entries 1-3, 5, 8) there
is a good agreement between the van der Waals radii,
rvdW, and the experimental hydrodynamic radii, rH. The
deviations [∆r/r in Table 1, where ∆r/r ) (rH - rVdW)/
rVdW] are smaller than 4% and comparable to the
experimental error of the measurements (Experimental
Section). In constrast, the experimental hydrodynamic
radii for protic solvents (Table 1, entries 4, 6, 7) are 20-
25% larger than the calculated van der Waals radii. This
reasonably reflects the self-association of alcoholic
solvents. Iterative calculations were carried out using
the initially derived hydrodynamic radius rH value as
rsolv in the Stokes-Einstein equation, eq 3, and compar-
ing the new rH value with the old rH. This process was
repeated several times until there was a convergence
between the last estimated hydrodynamic radius and
the preceding one introduced in eq 3. The resulting
hydrodynamic radii (indicated as rHcïnv in Table 1) were
used to determine the volumes of the diffusing species
in protic solvents.
(b) Determination of the Hydrodynamic Volume
of TMSS in Different Solvents. To check the reli-
ability of using rvdW and rHconv values in eq 3 for aprotic
and protic solvents, respectively, PGSE measurements
were performed for 0.1 mM solutions of TMSS [tetrakis-
(trimethylsilyl)silane], whose van der Waals radius is
known from the literature (rvdW ) 4.28 Å).20 The
experimental rH and c values are reported in Table 2.
Hydrodynamic radii that were derived by setting c equal
to 4 or 6 in eq 1 are also reported in Table 2 for
comparison. While there is a good agreement between
the estimated hydrodynamic radius of TMSS and its van
der Waals radius when the proper c factor is used (error
< 4%), the use of 6 or 4 in eq 1 leads to errors in the
5-30% range.
The effect of the concentration on viscosity and,
consequently, on the derived rH was checked by per-
forming PGSE NMR measurements for solutions of
TMSS (1 mM) in methylene chloride-d2 containing
different amounts of 3BPh4. Hydrodynamic radii of
TMSS determined by introducing the viscosity of pure
solvent in eq 3 or by taking into account the solution
viscosity are reported in Table 3. The solution viscosity
was calculated by multiplying the viscosity of the pure
solvent by a correction factor. The latter was equal to
the ratio of the slopes of the straight lines derived by
plotting log(I/I0) vs G2 for the resonances relative to
nondeuterated solvent residue in pure solvent and in
the solution.
(17) Handbook of Chemistry and Physics, 67th ed; CRC Press: Boca
Raton, FL, 1987.
(18) Holz, M.; Xi-an, Mao; Seiferling, D.; Sacco, A. J. Chem. Phys.
1996, 104, 669.
(20) Dinnebier, R. E.; Dollase, W. A.; Helluy, X.; Ku¨mmerlen, J.;
Sebald, A.; Schmidt, M. U.; Pagola, S.; Stephens, P. W.; van Smaalen,
S. Acta Crystallogr. 1999, B55, 1014.
(19) Mills, R. J. Phys. Chem. 1973, 77, 685.

Half-Sandwich Diamino Ru(II) Salts
Organometallics, Vol. 24, No. 14, 2005 3479
Table 2. Diffusion Coefficients (109Dt m2 s-1), c Factor Values, and Hydrodynamic Radii (rH) Obtained with
Different c Factors with Their Respective Percentage Errors, for Solutions of TMSS (0.1 mM) as a
Function of Solvent (Er at 25 °C)
Dt
c
rH(c)(∆r/r%)
rH(c)4)(∆r/r%)
rH(c)6)(∆r/r%)
1
2
3
4
5
6
7
8
benzene-d6 (2.27)
chloroform-d (4.81a)
CD2Cl2 (8.93)
2-propanol-d8 (19.92)
acetone-d6 (20.56)
ethanol-d6 (24.55)
methanol-d4 (32.66)
DMSO-d6 (46.45)
0.99
1.07
1.47
0.28
1.85
0.54
1.04
0.29
4.7
4.9
5.0
4.4
4.9
4.5
4.7
4.9
4.10 (4.2)
4.27 (0.3)
4.22 (1.5)
4.23 (1.2)
4.21 (1.7)
4.25 (0.8)
4.24 (0.9)
4.28 (0.1)
4.75 (10)
5.21 (21)
5.20 (21)
4.70 (10)
5.20 (21)
4.76 (11)
5.00 (16)
5.21 (21)
3.17 (26)
3.47 (19)
3.47 (19)
3.13 (27)
3.46 (19)
3.17 (26)
3.34 (22)
3.47 (19)
Table 3. Diffusion Coefficients (109Dt m2 s-1), c
Factor Values, and Hydrodynamic Radii (rH)
Obtained with Solvent Viscosity and Solution
Viscosity, for Solutions of TMSS (1 mM) in
Methylene Chloride-d2 at Various Concentrations
(C, mM) of 3BPh4
cationic or anionic moieties and that of the ion pairs,
N+ and N-, respectively, are reported in Table 4. They
represent a sort of aggregation number. Of course, a
distribution of ionic species is present in solution;
consequently N+ and N- indicate which is the apparent
average aggregation number of the ionic moieties. For
example, if they are both equal to 1 or 2, this means
that either ion pairs or ion quadruples, i.e., “(Ru+X-)2”,
are the predominant species in solution, respectively.
Less intuitive is the interpretation of N values when
odd aggregates are significantly present in solution, i.e.,
when N < 1 or N+ * N-, because of the different
volumes of the single ionic fragments. In our cases, they
are6,16e N+ ) 0.44 and N- ) 0.56 for 1BPh4, N+ ) 0.78
and N- ) 0.22 for 1PF6, N+ ) 0.47 and N- ) 0.53 for
2BPh4, N+ ) 0.81 and N- ) 0.19 for 2PF6, N+ ) 0.50
and N- ) 0.50 for 3BPh4, and N+ ) 0.83 and N- ) 0.17
for 3PF6. If N+ and/or N- are larger than the previously
indicated aggregation number of free ions, then ion
pairing and/or higher aggregation occurs.
solution viscosity
rH(c)(∆r/r%)
solvent viscosity
rH(c)(∆r/r%) rH(c)6)(∆r/r%)
C
Dt
c
c
1
3
4
5
6
0.7 1.39 5.0 4.28 (-0.01) 5.1 4.36 (1.8)
3.69 (-14)
3.82 (-10)
3.80 (-11)
4.06 (5)
12
24
1.34 5.1 4.29 (0.2)
5.1 4.47 (4.4)
1.35 5.1 4.28 (-0.01) 5.1 4.45 (4.0)
56
1.26 5.1 4.29 (0.2)
1.05 5.1 4.32 (0.9)
5.2 4.69 (9.4)
5.4 5.43 (27)
156
4.96 (15)
From the results reported in Table 3, it can be noted
that the hydrodynamic radius of the TMSS is nicely
reproduced in all cases if we use the solution viscosity
in eq 3. On the contrary, if the increased solution
viscosity derived from the addition of the 3BPh4 salt was
not taken into account, rH would be overestimated and
an apparent increased level of aggregation would be
deduced. Clearly, the error increases in cases of more
concentrated solutions. It is worth noting that if eq 1
with c ) 6 is used, the two errors derived from
considering (a) solvent instead of solution viscosity and
(b) 6 instead of the correctly determined c factor (which
tends to underestimate the hydrodynamic radius) go in
opposite directions and can compensate in some cases.
For example, for a 56 mM solution of 3BPh4 in meth-
ylene chloride-d2 (Table 3, entry 5), rH of TMSS is nicely
reproduced (error of 0.2%) if the solution viscosity and
the correct c factor are introduced in eq 3. The utiliza-
tion of the solvent viscosity and the correct c factor
affords a much higher error (9.4%), but if the solvent
viscosity and the c factor corresponding to the stick
boundary condition (c ) 6) are introduced in eq 3, the
two errors partially compensate (5%).
PGSE Measurements for Arene Ruthenium Com-
plexes Bearing Diamine Complexes. 1H and 19F-
PGSE NMR experiments were carried out for complexes
1, 2, and 3 in different solvents, with a relative permit-
tivity (ꢀr) ranging from 4.81 (chloroform-d) to 32.66
(methanol-d4), using TMSS as internal standard. PGSE
measurements allowed the translational self-diffusion
coefficients (Dt) for both cationic (Dt+) and anionic (Dt-)
moieties (Table 4) to be determined. By applying eq 3,
the average hydrodynamic radii and c factors for the
anionic and cationic moieties were measured (Table 4).
Their volumes were obtained from the average hydro-
dynamic radii of the aggregates, which were assumed
to be spherical. The volumes of the diffusing particles
were then compared with the van der Waals volumes
of ion pairs known from the solid state (Experimental
Section). The ratios between the apparent volume of the
All measurements indicated that a certain level of
aggregation was always present and that it was strongly
affected by the choice of ligand, solvent, and concentra-
tion. In constrast, the level of aggregation was not very
sensitive to the nature of the counterion.
(a) Ligand Effect on Aggregation: The Role of
the NH Group. The aggregation tendency of 1PF6,
2PF6, and 3PF6 was studied in methylene chloride-d2
at the same concentration (4 mM, entries 1, 4, 12 of
Table 4). The aggregation tendency follows in the order
1 > 2 ≈ 3. Compound 1PF6, which contains two NH2
functionalities, prevalently forms ion quadruples in
solution, while the introduction of two methyl groups
(2PF6) lowers the aggregation number, from 1.9 to 1.5
for N+ and from 1.9 to 1.4 for N-. Compound 3PF6,
which contains two NMe2 functionalities, shows less
aggregation tendency than 1PF6, but, interestingly
enough, it has the same N+ and similar N- as compound
2PF6. The difference between N+ and N- increases on
passing from NH2 functionalities to NMe2, indicating
that the presence of methyl tends to slacken the anion-
cation adducts. Another important comparison is the
trend of the aggregation number in methylene chloride-
d2 in the saturated solutions (entries 1, 5, 25). N+ and
N- are both equal to 1.9 and 1.5 for 1PF6 and 2PF6,
respectively, while they are equal to 2.0 and 1.9 for
3PF6. Probably the presence of methyl groups not only
induces a broadening of anion-cation adducts that
disfavors the formation of aggregates (1PF6 vs 2PF6) but
tends to increase the solubility of the adducts. The
increased solubility for 3PF6 permits increasing the
concentration by about 30 times; this could force the ion-
pairs/ion-quadruples equilibrium to shift to the right

3480 Organometallics, Vol. 24, No. 14, 2005
Zuccaccia and Macchioni
Table 4. Diffusion Coefficients (1010Dt m2 s-1), Hydrodynamic Radii (rH, Å), c Factors, and Aggregation
Numbers (N) for Compounds 1, 2, and 3 as a Function of Solvent (Er at 25 °C) and Concentration (C, mM)
+
+
-
-
Dt+
Dt-
rH
cH
rH
cH
N+
N-
C
1PF6 (Vip ) 283)
CD2Cl2 (8.93)
2-propanol-d8 (19.92)
1BPh4 (Vip ) 507)
CD2Cl2 (8.93)
1
2
11.0
2.0
10.9
2.6
5.1
5.0
5.3
4.9
5.0
4.3
5.3
4.5
1.9
1.9
1.9
1.1
4
4
3
8.5
8.8
6.2
5.6
6.1
5.5
2.0
1.9
4
2PF6 (Vip ) 315)
CD2Cl2 (8.93)
CD2Cl2 (8.93)
4
5
10.9
11.6
11.4
11.4
4.8
4.8
5.3
5.3
4.7
4.8
5.2
5.3
1.5
1.5
1.4
1.5
4
6b
2BPh4 (Vip ) 540)
CD2Cl2 (8.93)
6
7
8
9.1
8.6
8.3
8.9
8.6
8.3
5.7
5.8
5.9
5.5
5.5
5.5
5.8
5.8
5.9
5.5
5.5
5.5
1.4
1.5
1.6
1.5
1.5
1.6
4
CD2Cl2 (8.93)
12
CD2Cl2 (8.93)
15b
3PF6 (Vip ) 347)
chloroform-d (4.81a)
CD2Cl2 (8.93)
9
7.5
11.3
10.9
11.2
9.8
7.6
11.9
11.7
11.6
10.7
9.9
5.5
4.8
4.9
4.9
5.1
5.2
5.4
5.6
4.8
5.0
5.3
5.2
5.2
5.3
5.4
5.4
5.4
5.4
5.2
5.1
5.4
4.7
4.7
4.8
4.8
5.1
5.1
5.4
3.6
3.9
5.3
5.2
5.2
5.3
5.2
5.3
5.3
5.4
3.7
4.5
2.0
1.3
1.4
1.5
1.7
1.8
1.9
2.1
1.3
1.5
1.9
1.1
1.2
1.3
1.4
1.6
1.7
2.0
0.5
0.7
1.15b
0.7
2.7
4
10
11
12
13
14
15
16
17
18
CD2Cl2 (8.93)
CD2Cl2 (8.93)
CD2Cl2 (8.93)
18
CD2Cl2 (8.93)
9.4
34
80
CD2Cl2 (8.93)
8.6
8.4
16.2
8.8
9.2
CD2Cl2 (8.93)
8.7
25.3
12.9
115b
4
acetone-d6 (20.56)
methanol-d4 (32.66)
3BPh4 (Vip ) 571)
chloroform-d (4.81a)
CD2Cl2 (8.93)
6b
19
20
21
22
23
24
25
26
6.3
10.7
9.9
6.3
9.8
9.2
8.6
8.3
7.6
6.0
14.8
6.4
5.1
5.4
5.6
5.8
6.0
6.3
5.0
5.5
5.4
5.4
5.4
5.5
5.5
5.6
5.2
6.4
5.3
5.7
5.9
6.0
6.1
6.4
5.0
5.5
5.3
5.5
5.5
5.5
5.5
5.6
5.2
1.9
1.0
1.2
1.3
1.4
1.6
1.9
0.9
1.9
1.1
1.4
1.5
1.6
1.7
2.0
0.9
0.5b
0.7
2.7
12
24
56
CD2Cl2 (8.93)
CD2Cl2 (8.93)
9.1
CD2Cl2 (8.93)
8.6
CD2Cl2 (8.93)
7.8
CD2Cl2 (8.93)
6.1
156
4
acetone-d6 (20.56)
14.7
a 20 °C. bSaturated solution.
Scheme 3
(X ) BPh4 and PF6) in methylene chloride-d2. Data are
reported in Table 4, entries 10-16 for 3BPh4 and entries
20-25 for 3PF6. As expected, the aggregation increased
with the concentration, but it is significant even at the
lowest concentration considered (0.7 mM), for which the
aggregation numbers were higher than 1 for both
complexes.
Intramolecular Characterization in Solution:
Conformational Analysis of Cymene Rotation
through Intramolecular NOE NMR Experiments.
All complexes 1-3 were characterized in solution by 1H,
13C, and 19F NMR spectroscopies. Data are reported in
the Experimental Section. Numbering of carbon and
proton resonances is illustrated in Scheme 3.
The higher symmetry of 1 and 3 complexes makes the
normal and prime of cymene protons and the two 8H′/
9H′ and 8H/9H couples magnetically equivalent. This
reduces the number of observed NMR signals and also
the availability of spatial “reporters” in the cationic
moiety. As a consequence, the results related to com-
plexes 2 will have a dominant position in the following
considerations on arene orientations and relative anion-
cation positions.
Figure 2. Trends of the aggregation numbers (N+ and N-)
of compound 3PF6 as a function of the relative permittivity
of the solvent.
even if ion quadruples are energetically less favored
than in 1,2PF6.
(b) Solvent Effect on Aggregation. The aggrega-
tion tendency of 3PF6 was studied in several solvents.
The dependence of N+ and N- on ꢀr is reported in Figure
2. It clearly shows that N+ and N- increase as ꢀr
decreases, reaching the maximum value of 2 for chlo-
roform-d where quadrupoles are mainly present. The
minimum level of aggregation is observed in acetone-
d6, where still a significant amount of ion pairs is
present. Another clear trend shown in Figure 2 is that
the N+ value is higher than that of N- with a difference
that increases with ꢀr up to acetone-d6. Surprisingly, a
comparison between acetone-d6 (ꢀr ) 20.56) and meth-
anol-d4 (ꢀr ) 32.66) indicates that the aggregation
tendency is higher in the latter.
(c) Concentration Effect on Aggregation. The
effect of concentration was investigated for complex 3X

Half-Sandwich Diamino Ru(II) Salts
Organometallics, Vol. 24, No. 14, 2005 3481
The latter criterion for determining the preferential
orientation of the cymene (based on the differential Meu
and Med/2, 2′, 3, 3′ NOE intensities) is strengthened by
the observations that the Hu proton affords a strong
NOE with 5-Me cymene protons, while no interaction
is present with 6 and 7 (Supporting Information).
Once the R resonances bonded to N have been
assigned, it is then possible to assign the four methylene
protons due to their selective dipolar interactions; the
proton labeled 8′ interacts only with Hu proton and
proton 9 interacts only with Hd (Supporting Informa-
tion). Finally, the 1H,13C-HMQC NMR spectrum shows
the scalar coupling of pair 8 and 8′ (and also 9 and 9′)
with the same carbon, and so the four methylene
protons are assigned (Supporting Information).
(b) Complex 2PF6. The selective dipolar interactions
of R groups with cymene protons for this complex also
indicate that the conformation with the methyl group
oriented toward the Hu proton is more populated than
the others, analogous to 2BPh4.
1
Figure 3. Section of H-NOESY NMR spectrum (400.13
MHz, 296 K, methylene chloride-d2) of complex 2BPh4
showing the different interactions of Meu and Med with
cymene protons.
(c) Complexes 1X and 3X. In complexes 1 and 3, 2
and 3 cymene protons are magnetically equivalent
(Scheme 3). Consequently, information about the con-
From the 1H, 13C, 1H-COSY, 1H-NOESY, 1H,13C-
1
HMQC NMR, and H,13C-HMBC NMR spectroscopies
1
it is rather easy to group the proton and carbon
resonances belonging to the different fragments. On the
contrary, the distinction between the resonances of (1)
the methylene protons, (2) the R groups bonded to the
nitrogen, and (3) the cymene ligands and the determi-
nation of the preferred conformers of cymene are two
strictly interlocked and difficult issues that, neverthe-
less, can be settled by means of 1H-NOESY experiments,
as detailed in the following sections.
formation of the cymene is difficult to obtain from H-
NOESY NMR spectra. Selective dipolar interactions are
observed between the protons of Ru and the protons of
cymene, while very weak (R ) Me) or no (R ) H) NOEs
are present between the latter and the Rd protons. For
compound 1, the presence of stronger interaction be-
tween Hu and 5 with respect to Hu and 2 or 3 suggests
that in the preferential orientation of the cymene 5-Me
is directed toward the nitrogen atom bearing proton Hu.
For compound 3, the NOE between Meu and 3 is twice
as intense as that between Meu and 2. This suggests
that 5-Me is directed toward the nitrogen atom (as we
have seen for complexes 2 and 3) but is in a more central
position with respect to the two N arms, probably due
to the steric hindrance of the methyl groups. Once the
two different resonances of the R groups are assigned,
the two methylene protons (normal from prime) can be
distinguished through considerations similar to those
made for complex 2.
(a) Complex 2BPh4. As reported in Figure 3, the 1H-
NOESY spectrum shows that the dipolar interactions
between the cymene protons and one N-Meu group
(labeled u just because it points up toward the cymene)
have higher intensities than the analogous one of the
other methyl Med protons (d stands for down). The two
Hd and Hu broad resonances, which are scalar coupled
with the methyl Meu and Med, respectively, also show
different dipolar interactions with cymene protons. Only
the NOEs between the Hu and cymene resonances are
visible. An in-depth analysis of the section of 1H-NOESY
spectrum reported in Figure 3 shows that the intensity
of the NOEs between Meu and cymene protons follows
the order 2 ≈ 3 > 2′ . 3′. In addition, Med gives a
medium-size NOE with 3′ and a weak NOE with 2′.
These observations are consistent with the main
presence of a preferential orientation of the cymene in
which 5-Me is directed toward the nitrogen atom bear-
ing proton Hu with the 7-iPr group located in the
nonhindered region between the other nitrogen atom
and chlorine (Figure 3). On the other hand, this cymene
orientation does not explain the NOE between Meu and
2′, which has a higher intensity than those of 6 with 2
and 2′. The other orientation with the 5-Me directed
toward the chlorine (Figure 3) must be present in
solution, even in a limited percentage, to maintain the
observed specificity. In Figure 3 the two cymene orien-
tations are reported as eclipsed because such a situation
has been observed for compound 1PF6 in the solid
state.16e However, our NOE analysis does not indicate
for certain if the preferred orientations are eclipsed or
staggered.
Interionic NOE NMR Measurements. The relative
anion-cation orientations in solution for complexes
1-3X were studied by detecting dipolar interionic
interactions in the 19F,1H-HOESY (X- ) PF6-) and 1H-
NOESY (X- ) BPh4-) NMR spectra at room tempera-
ture (296 K). The following solvents, differing in both
relative permittivity and nature, were taken into ac-
count: methylene chloride-d2 (ꢀr25°C ) 8.93), 2-propanol-
25°C
25°C
d8 (ꢀr
) 19.92), and methanol-d4 (ꢀr
) 32.66). In
some cases, after having evaluated the longitudinal
relaxation times (Experimental Section), quantitative
NOE experiments (mixing time ) 150 ms, recycle time
) 10 s) were carried out and the derived data were
treated taking into account that the volumes of the NOE
cross-peaks are proportional to (nInS/nI + nS) where nI
and nS are the number of equivalent I and S nuclei,
respectively (Table 5).21 The preferred relative anion-
cation orientations that were deduced from the observed
interionic NOEs depended very little on solvent, while
(21) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95.

3482 Organometallics, Vol. 24, No. 14, 2005
Zuccaccia and Macchioni
-
Table 5. Relative NOE Intensities Determined by
Arbitrarily Fixing at 1 the Intensity of the NOE(s)
between the Anion Resonances (o-H in the Case of
BPh4-) and the Methyl (5) of Cymene
indicate that PF6 is located above the two methylene
carbons and the two amine nitrogen moieties and it is
shifted toward the less hindered N-arm having the
R-group pointing toward the chlorine atom. From this
position, the anion can interact with the “prime” cymene
protons oriented far away from the chlorine. The
detected dipolar interaction with 3′ can be explained by
the presence of the other cymene orientation described
in the previous section.
a
2BPh4
3PF6
a
a
1PF6
o
m
2PF6
a
b
2′
2.7
1.7
0.4
0.8
1.4
1
0.2
0.2
0.2
0.3
0.4
0.1
0
1
0.9
2
1.1
3
3.0
1.7
1.4
3′
c
1
0.5
0
2.2
0
In the cases of complexes 1 and 3, the location of the
counterion was necessarily less defined because of the
magnetic equivalence of the two sides of both cymene
and the N,N ligand. Despite this, for compound 1,
selective dipolar interactions were observed as far as
the diamine resonances are concerned; that is, only the
prime or up protons showed NOEs with the counterion.
The 2 and 3 aromatic cymene resonances interacted
dipolarly with the counterion with the same strength
as PF6-/Hu in 1PF6 (3.0 vs 2.7 and 3.2 in Table 5), while
in compound 2, the former interactions were less intense
(1.1 vs 2.5 in Table 5). On the contrary, for complex 3
nonnegligible interionic contacts were observed between
the counterion and the methylene protons that point far
away from the cymene (H8) and Med (Table 5). As a
consequence, it can be concluded that in complex 1 the
anion is located in a more central position with respect
to the N,N ligand than in 2 because 1 has equally
hindered N-arms; nevertheless, the anion is still located
above the two methylene carbons and the two amine
nitrogen moieties (CCNN) that are closer to the aro-
matic cymene resonance. In complex 3, the anion is also
located in a central position, but the presence of methyl
groups on both nitrogen atoms probably moves the anion
away from the cymene and also makes the position
below the CCNN moiety accessible.
5
1
0.4
1
0.4
1
0.6
7
0.1
7′
8′
1.2
0.4
0.3
0.3
0.1
1.1
0.7
1.4
1
8
9′
1.3
0
9
0.4
0.3
0.2
0
0.2
0.5
0.2
0
Hu
Meu
Hd
Med
3.2
0
2.5d
0.3
0
1.6
0.5
1.5
0.9
0.4
0.2
1.2
a In methylene chloride-d2 at 296 K. b In methanol-d4 at 296
K. e Difficult to quantitatively evaluate due to overlap with Hu.
d Probably overestimated due to the superimposition of 3′.
(b) X- ) BPh4-. For this anion, much stronger NOEs
were observed between the anion and all cymene and
methylene protons. As can be seen from Table 5, the
NOE intensities of 8 and 9 with o-H were comparable
to or even higher than those of 8′ and 9′. The same
situation was found for cymene resonances (compare 2′
and 3′ with respect to the 2 and 3). It is possible that
the voluminous BPh4- anion cannot approach the small
cavity formed by the cymene and N,N ligand. As a
consequence, the anion can locate on the side of the N,N
ligand or above cymene. The energies of the two relative
anion-cation orientations may differ very little. Another
possibility is that the counterion may assume an
intermediate position with one phenyl above the cymene
and another directed toward the ligand. Finally, the
interionic contacts of o-H and m-H resonances and
cationic protons have completely different intensity
trends. NOEs between m-H and cationic protons are less
intense that those of o-H, but the intensity depends very
little on the type of cationic protons (Table 5), while o-H
shows NOE intensities that are strongly affected by the
type of cationic protons.
Figure 4. Sections of 19F-1H-HOESY NMR spectrum
(376.65 MHz, 296 K, methylene chloride-d2) of complex
2PF6. * denotes the resonance due to a residue of non-
deuterated solvent. ** denotes the resonance of H2O.
they were remarkably affected by the nature of the
counterion and the R groups bonded to the nitrogen
atoms.
(a) X- ) PF6-. For complex 2, strong NOE contacts
were observed between F atoms of the counterion and
8′, 9′, Hu, and Med resonance (Table 5 and Figure 4).
Medium-size NOEs were also detected with 2-3, 3′, 5
cymene resonances, while weak NOEs were observed
with 7 of the isopropyl group of cymene and with Med.
The anion did not show any interaction with 2′, 7′, 8,
and 9 protons. It is interesting to notice that the anion
“sees” all the cymene protons with the exception of 7′
and 2′ (see the expansions in Figure 4). As for the anion
interactions with diamine protons, the protons labeled
as up or prime “see” the anion, while among the others
only Med interacts with it. All these observations
Discussion
Methodology for Obtaining Accurate rH and VH
Values by PGSE NMR Measurements. The results
of PGSE NMR measurements carried out on pure sol-
vents (Table 1), TMSS solutions in various solvents
(Table 2), and TMSS in methylene chloride-d2 at vari-

Half-Sandwich Diamino Ru(II) Salts
Organometallics, Vol. 24, No. 14, 2005 3483
able concentrations of 3BPh4 (Table 3) indicate that
accurate rH and VH values can be obtained for medium-
size molecules by treating the Dt data with eq 3. Our
findings show that the utilization of the van der Waals
radius instead of the hydrodynamic radius of solvent
(rsolv) in eq 3 is a good approximation only for aprotic
solvents. For protic ones, the rsolv to be introduced in eq
3 can be derived through an iterative process starting
from rsolv ) rVdW and then using the derived hydrody-
namic radius, rH, value as rsolv until there is a conver-
gence between the last two derived rH values. By using
the latter hydrodynamic radius (rHconv) as rsolv in eq 3
for protic solvents, there is good agreement between the
found rH and the van der Waals radius of TMSS (entries
4, 6, and 7 in Table 2). The viscosity value to be
introduced in eq 3 has to be corrected by multiplying
the viscosity value of pure deuterated solvents, at the
proper temperature, by a factor that is the ratio of the
diffusion coefficients of an internal standard (TMSS or
the solvent itself) in the solution and in the pure solvent.
Since the reliability of eq 3 has been checked, we here
propose a simplified methodology for obtaining accurate
rH values by performing PGSE measurements of the
sample of interest containing an internal standard (that
can also be the solvent itself). From eq 1, the ratio of
the Dt coefficients for the two species is equal to
Figure 5. [1 × 1] network of hydrogen bonds between Cl
and NH in 1PF6 that connects two cationic Ru(II) moieties
in the solid state (from ref 16e). Interionic NH‚‚‚FPF5- HBs
are also shown.
4). Since the tendency to afford intercationic HBs of 1
and 2 should be similar, this different behavior could
be due to the nucleation process that, in the case of 1,
could go through the association of ion quadruples, while
in the case of 2, could involve an ion quadruple and an
ion pair. Even though this issue is highly speculative,
simple modeling processes indicate that two ion qua-
druples can favorably interact in 1 using the N-H
moieties not involved in the formation of ion quadruples,
while in 2, this is not possible because of the presence
of the N-Me groups. To obtain information about the
relative force of the HB networks that hold the two
cations and the cation and the anion together, PGSE
measurements in 2-propanol-d8 were performed for
1PF6 (entry 2 in Table 4). The results clearly indicate
+
-
that ion triples 12PF6 and PF6 are the predominant
Dsta cst rsHt
Dstt csa rsHa
species in solution. The theoretical N+ and N- values
)
(4)
+
-
for a solution containing only 12PF6 and PF6 are 1.8
and 1.0, respectively; these values are very similar the
observed ones (N+ ) 1.9 and N- ) 1.1, entry 2 in Table
4). Since the pattern of interionic NOEs is not different
from that in other aprotic solvents, it is reasonable to
assume that the observed ion triples have the “PF6-
RuRu+” structure; that is, they derive from the dissocia-
tion of an anion from the ion quadruples without
destroying their structures. This means that for an ion
quadruple in a solvent such as 2-propanol-d8, that is
suitable to solvate anions,22 it is preferable to dissociate
an anion rather than form two ion pairs. In agreement
with these results, neutral catalysts for transfer hydro-
genation in 2-propanol-d8 exist mainly as dimers and
the tendency to form dimers is higher than that of
compounds 1 and 2.13
where sa and st stand for sample and standard, respec-
tively. According to eq 2, c ) f(rsolv, rH). Since the Dt
ratio is measured and rsolv and rsHt are known, or can be
measured as illustrated above, the only unknown pa-
rameter in eq 4 is rsHa, which can be derived without
having to take into account the solution viscosity or
temperature. In addition, if the absolute Dt value of the
sample is not needed, it is not even necessary to
calibrate the gradient strength.
Interionic Structure of Diamino Ru(II) Salts.
PGSE NMR investigations reveal that complexes 1-3
have a marked tendency to form not only ion pairs but
also higher aggregates (up to ion quadruples). The main
factor that determines aggregation is the possibility of
N-H groups of the diamine ligand to form hydrogen
The lower tendency of compound 3, which has no N-H
group, to afford ion pairs and ion quadruples is intrigu-
ing and in agreement with previous results.12 NOE
investigations show that anion-cation interactions are
weaker than in 1 and 2 and the anion now also interacts
with protons 8 and 9 and with Med. This indicates that
less intimate ion pairs are present in solution with a
consequent increased dipole moment. In apolar solvents
at high concentrations, these ion pairs with elevated
dipole moment may persist in solution by reducing their
dipole moment through the formation of ion quadruples
(entries 16 and 25 in Table 4). Finally, when the ꢀr of
aprotic solvents increases, the aggregation level of 3PF6
decreases, while the difference between N+ and N-
increases (Figure 2) in agreement with previous stud-
ies.12 Interestingly, in methanol-d4 the aggregation level
is higher than in the less polar acetone-d6 (entries 17,
18 in Table 4). In the former solvent the N+ and N-
-
bonds (HBs) with the PF6 anion and with the Ru-Cl
moiety of another cationic unity. Clearly, this is not the
only cause of aggregation since even when N-H groups
are not present (as in compound 3) and the anion is
BPh4-, ion pairs and ion quadruples may still form
(entries 19-25 in Table 4).
Compounds 1 and 2 can form ion quadruples through
a [1 × 1] network of intercationic HBs between Ru-Cl
and the N-H groups that point down as observed in the
solid state for 1PF6 (Figure 5).16e N-H moieties that
point up toward the cymene are committed in HBs with
the anion as observed in the solid state for 1PF6 (Figure
5) and in solution for 1PF6 and 2PF6 by means of our
interionic NOEs measurements. Interestingly, while ion
quadruples are the predominant species in saturated
solutions of compound 1 in apolar and aprotic solvents
(entries 1 and 3 of Table 4), in the same conditions,
compound 2 leads to equimolar solutions of ion pairs
and ion quadruples (N+ ≈ N- ≈ 1.5, entries 4-8 of Table
(22) Reichardt, C. Solvents and Solvent Effect in Organic Chemistry;
Wiley-VCH: Weinheim, 2003.

3484 Organometallics, Vol. 24, No. 14, 2005
Zuccaccia and Macchioni
values are 1.5 and 0.7, respectively, which suggests that
ion triples, analogous to those observed in 2-propanol-
d8, are significantly present. Ion triples in methanol-d4
-
have the anion PF6 in the usual position as deduced
by the observation of heteronuclear NOEs in the 19F,1H-
HOESY experiments. In this respect, it must be said
that anion-cation interionic NOEs were observed in all
solvents. The same NOEs were detected also with
similar relative intensities (see for example columns a
and b in Table 5 pertinent to compound 3PF6). For any
particular compound, the absolute intensity of NOEs
was higher in solvents where a higher percentage of
intimate aggregates was present.
Conformational Analysis of Cymene Rotation
through NOE NMR Experiments. The arene orien-
tation in half-sandwich arene complexes has been
investigated extensively in the solid state.23 The stag-
gered conformations are usually favored24-27 but the
eclipsed ones have been observed in some cases.28,29
In solution, restricted rotation about the arene metal
bond has rarely been observed.30 Theoretical investiga-
tions indicate that the rotational energetic barrier of the
arene ligand is usually low (10.6 kJ mol-1 in [(η6-C6H6)-
RuB(pz)3]PF623 and 15.5-19.7 KJ mol-1 for [η6-(C6H6)-
Cr(CO)3]31). This means that eclipsed, staggered, or
intermediate conformations are all accessible. NOE
experiments have been used to determine the arene
orientation in solution.16d,32-34 Sadler and co-workers
have described the behavior of the arene in the complex
[(η6-arene)RuY(en)]X where en ) ethylendiamine, arene
) biphenyl, tetrahydroanthracene, dihydroanthracene,
and Y ) 9EtG- N7 (guanine base coordinate with N7)
or their parent Cl.16d In all cases, except for the
compound with biphenyl and Cl, in the solid state the
arene orients the substituent toward the region between
the chlorine or guanine base and one nitrogen of the
ligand. A similar situation was found in solution by
means of NOESY experiments.
Figure 6. Four possible eclipsed (n) conformers of cymene
and their relative staggered conformations ((δ).
tween Cl and NR1R2. In solution, the cation structure
appears to be similar due to the NOE contacts between
Hu and 5-Me and the absence of the contact between
Hu and 7-iPr. However, this analysis does not allow the
orientations ranging from eclipsed to staggered in the
δ interval (1 in Figure 6) to be distinguished. A small
percentage of the eclipsed or staggered conformation
with methyl above the Cl (3 in Figure 6), in which the
isopropyl group is between the NH functionalities, must
be present. The other 2 and 4 conformations were not
observed, probably because they are less stable than the
1 and 3 due to the steric repulsion between bulky
isopropyl group and the ligand.
The two additional “anti” methyl groups in complex
2PF6 increase the number of observed NMR signals,
making more spatial reporters in the cationic moiety
available for investigating the arene conformation. Even
in this case, the preferential orientation of the cymene
is with the 5-Me directed toward the nitrogen atom
bearing proton (Hu) and, consequently, the 7-iPr group
is located in the nonhindered region between the other
nitrogen atom and chlorine (1 in Figure 6). The other
orientation with the 5-Me directed toward the chlorine
(3 in Figure 6) must still be present in solution even if
at a lower percentage in order to maintain the observed
specificity of NOE contacts. The conformation in which
the 5-Me is directed toward the nitrogen bearing Meup
is probably not present due to the steric hindrance of
the methyl.
Our NOE experiments for complexes 1, 2 and 3 show
that the conformation of the arene depends on the R
substituent of the nitrogen ligand.
In the solid state,16e complex 1PF6 shows an eclipsed
conformation with the cymene methyl close to the
nitrogen (1 in Figure 6); in this configuration, the bulky
isopropyl group occupies the free region of space be-
The complex 3PF6, where every N atom has two
methyl groups, the conformers 1 and 4 of Figure 6 are
not stable, probably due to the steric hindrance between
NMe2 and 5-Me or 7-iPr. Of the remaining conformers
2 and 3, the former is the preferred one, as indicated
by the NOE between Meu and protons 3 that is twice
as intense as that of the Meu-2 NOE.
In summary, NMR data indicate that arene orienta-
tion in complexes 1-3 is mainly dictated by steric effects
and, in particular, by the minimization of the steric
repulsion between 5-Me and 7-iPr groups and the
ligands. As a consequence, when protons are present
in the N-arms, as in complex 1, conformer 1 is the most
abundant, but a small percentage of 3 is present. For
complex 3, when the steric hindrance of N-arms is
considerable, 5-Me and 7-iPr are directed toward Cl and
conformer 2, having 7-iPr close to Cl, is favored.
These results are in perfect agreement with the
results reported by Sadler and co-workers;16d in fact,
(23) (a) Bhambri, S.; Bishop, A.; Kaltsoyannis, N.; Tocher, D. A. J.
Chem. Soc., Dalton Trans. 1998, 3379. (b) Albrigth, T. A. Acc. Chem.
Res. 1982, 15, 149.
(24) Bailey, M. F.; Dahl, L. F. Inorg. Chem. 1965, 4, 1298.
(25) Koshland, D. E.; Myers, S. E.; Chesick, J. P. Acta Crystallogr.,
Sect. B 1977, B33, 2013.
(26) Bennet, M. A.; Robertson, G. B.; Smith, A. K. J. Organomet.
Chem. 1972, 43, C41.
(27) Restivo, R. J.; Ferguson, G.; O’Sullivan, D. J.; Lalor, F. J. Inorg.
Chem. 1975, 12, 3046.
(28) (a) Iverson, D. J.; Hunter, G.; Blount, J. F.; Damewood, J. R.,
Jr.; Mislow, K. J. Am. Chem. Soc. 1981, 103, 6073. (b) Hunter, G.;
Blount, J. F.; Damewood, J. R.; Iverson, D. J.; Mislow, K. Organome-
tallics 1982, 1, 448.
(29) Sledle, A. R.; Newmark, R. A.; Pignolet, L. H.; Wang, D. X.;
Albright, T. A. Organometallics 1986, 5, 38.
(30) Albrigth, T. A.; Hoffmann R.; Tse J. C.; D’Ottavio T. J. Am.
Chem. Soc. 1979, 101, 3812.
(31) Braga, D. Chem. Rev. 1992, 92, 633.
(32) Wright, J. M.; Landis, C. R.; Ros, M. A. M. P.; Horton, A. D.
Organometallics 1998, 17, 5031.
(33) Landis, C. R.; Luck, L. L.; Wright, J. M. J. Magn. Reson., Ser.
B 1995, 109, 44-59.
(34) Beringhelli, T.; D’Alfonso, G.; Maggioni, D.; Mercandelli, P.;
Sironi, A. Chem. Eur. J. 2005, 11, 650.

Half-Sandwich Diamino Ru(II) Salts
Organometallics, Vol. 24, No. 14, 2005 3485
and 85% H3PO4 (31P). NMR samples were prepared by dis-
solving the suitable amount of compound in 0.5 mL of solvent.
Synthesis of Complexes 1-3X.15 Synthesis of complex
1BPh4. [Ru(η6-arene)Cl2]2 (0.100 g, 0.163 mmol) was added
to a solution of 24 µL (0.360 mmol) of the ethylenediamine in
5 mL of MeOH. The resulting red-orange suspension was
stirred for 3 h at rt until it changed into a yellow solution; a
large excess of NaBPh4 (10 equiv) dissolved in 0.5 mL of MeOH
was added, and a precipitate formed. The solution was filtered
and the solid was washed with cold MeOH and n-hexane. Yield
) 84%. 1H NMR (CD2Cl2, 298 K, 400.13 MHz, J in Hz): δ 1.26
even if the arene is different (in nature or in the position
of the substituents), in the chlorine derivatives, the
substituent on the arene is located between the Cl and
NH2 moiety in both solid state and solution.
Conclusions
From the methodological point of view, the results
reported in this paper show the importance of elaborat-
ing the Dt data deriving from PGSE NMR measure-
ments through the Stokes-Einstein equation in which
the proper c factor is introduced (eq 3). The latter,
expressed as a function of solvent radius and hydro-
dynamic radius of the diffusing particles, differs con-
siderable from both 4 (slip boundary conditions) and 6
(stick boundary conditions) for medium-size molecules
usually used in organometallic chemistry. Consequently,
only by using an appropriate c factor can the rH and VH
values be accurately determined.
The refined PGSE methodology combined with NOE
NMR experiments affords a clear picture of the
interionic structure of compounds 1-3, which exhibit a
remarkable tendency to aggregate, not limited to ion
pairing, in a variety of solvents. In fact, for compounds
1 and 2, ion quadruples that are held together by a
[1 × 1] network of intercationic HBs between Ru-Cl and
the N-H groups are formed in solution. Interestingly
enough, in 2-propanol-d8, ion quadruples prefer to
dissociate into an anion and a “XRuRu+” ion triple
rather than into two ion pairs. Compelling evidence for
this process comes from the aggregation numbers, N+
and N-, determined by PGSE measurements, that agree
with those expected for ion triples. Another clue about
the “RuRu” approach is derived from interionic NOE
experiments that indicate that relative anion-cation
orientation is the same in ion pairs, ion triples, and ion
quadruples.
Compound 3, which has no N-H moiety, has a lesser
tendency to aggregate than 1 and 2, even though it also
affords ion quadruples in apolar and aprotic solvents
probably because it has to reduce the molecular dipole
moment in order to stay in solution. In 3PF6 the anion
location is less specific than in 1PF6 and 2PF6 as a
consequence of the less intimate ion pair due to the
absence of NH‚‚‚FPF5 HBs.
Finally, the quantitative evaluation of intramolecular
NOEs indicates that particular conformations of cymene
in 1-3 are more abundant than others in solution, and
this is mainly determined by steric factors.
3
(d, JH7-H6 ) 6.95, H7), 1.72 (m, H8), 2.01 (m, H8), 2,04 (s,
H5), 2.12 (br, Hd), 2.62 (sept, 3JH6-H7)7.19 H6), 3.09 (br, Hu),
3
3
4.98 (d, JH3-H2 ) 5.95 H3), 5.17 (d, JH2-H3 ) 6.00, H2), 6.96
3
3
(t, Jp-m ) 7.19, p), 7.10 (t, Jm-o,p ) 7.36, m), 7.43 (br, o).
Synthesis of Complex 1PF6. Pathway a. [Ru(η6-arene)-
Cl2]2 (0.100 g, 0.163 mmol) was added to a solution of 24 µL
(0.360 mmol) of the ethylenediamine in 5 mL of MeOH. The
resulting red-orange suspension was stirred for 3 h at rt until
it changed into a yellow solution; a large excess of NH4PF6
(10 equiv) dissolved in 0.5 mL of MeOH was added, and a
precipitate formed. The solution was filtered, and the solid was
washed with cold MeOH and n-hexane. Yield ) 80%. Pathway
b. 1BPh4 (0.100 g, 0.154 mmol) was dissolved in 5 mL of CH2-
Cl2. A 0.056 g sample of TlPF6 (0.161 mmol) was added under
nitrogen atmosphere, and TlBPh4 precipitated from the solu-
tion. The solution was filtered and dried under vacuum, giving
1
a yellow solid. Yield ) 98%. H NMR (CD2Cl2, 298 K, 400.13
3
MHz, J in Hz): δ 1.34 (d, JH7-H6 ) 6.94, H7), 2.28 (s, H5),
2.69 (m, H8 e 8′), 2.87 (sept, 3JH6-H7 ) 6.83 H6), 3.11 (br, Hd),
3
3
4.91 (br, Hu), 5.52 (d, JH3-H2 ) 5.94 H3), 5.67 (d, JH2-H3
)
6.00, H2). 19F NMR (CD2Cl2, 298 K, 376.65, J in Hz): δ -75.10
1
-
(d, JFP ) 707.9, PF6
)
Synthesis of Complex 2,3X (X ) BPh4, PF6). Complexes
2,3X (X ) BPh4, PF6) were synthesized with the same
procedures as 1X using the appropriate ligand. Yields were
in the 85-95% range. 2BPh4: 1H NMR (CD2Cl2, 298 K, 400.13
MHz, J in Hz): δ 1.28 (d, 3JH7′-H6 ) 7.1, H7′), 1.30 (d, 3JH7-H6
) 7.1, H7), 1.85 (m, H9′), 1.95 (m, H8), 2.04 (s, H5), 2.14 (m,
H8′), 2.40 (m, H9), 2.61 (d, 3JMe“d”-H“u” ) 5.76, Med), 2.80 (sept,
3
3JH6-H7,H7′ ) 7.0, H6), 2.87 (d, JMe“u”-H“d” ) 6.19, Meu), 3.45
3
(br, Hd), 4.04 (br, Hu), 5.02 (d, JH3′-H2′ ) 6.09, H3′), 5.06 (d,
3
3JH3-H2 ) 6.04, H3), 5.14 (d, JH2-H3 ) 2.74, H2), 5.21 (d,
3JH2′-H3′ ) 5.93, H2′), 6.96 (t, 3Jp-m ) 7.19, p), 7.10(t, 3Jm-o,p
)
7.36, m), 7.43 (br, o). 13C{1H} NMR (CD2Cl2, 298 K, 100.55
MHz): δ 18.4 (s, C5), 22.1 (s, C7 or C7′), 22.4 (s, C7′ or C7),
31.3 (s, C6), 44.3 (s, Med), 45.7 (s, Meu), 53.2 (s, C8), 57.1 (s,
C9), 81.3 (s, C2), 82.1 (s, C2′), 82.7 (s, C3), 82.8 (s, C3′), 86.0
3
(s, C1), 106.1 (s, C4), 122.4 (s, p), 126.1 (q, Jm-B ) 2.7, m),
136.3 (d, 2Jo-B ) 1.4, o), 164.4 (q, 1JC-B ) 49.2, C-ipso). 2PF6:
1H NMR (CD2Cl2, 298 K, 400.13 MHz, J in Hz): δ 1.33 (d,
3JH7′-H6 ) 6.92, H7′), 1.34 (d, 3JH7-H6 ) 6.90, H7), 2.20 (m, H8),
2.34 (s, H5), 2.53 (m, H9′), 2.73 (m, H8′ e H9), 2.88 (d,
3JMe“d”-H“u” ) 5.78, Med), 2.98 (sept, 3JH6-H7,H7′ ) 7.0, H6), 3.14
3
(d, J
) 6.21, Meu), 3.61 (br, Hd), 5.41 (m, H2 e H3),
Me“u”-H”d”
3
3
5.46 (d, JH3′-H2′ ) 6.16, H3′), 5.55 (d, JH2′-H3′ ) 6.03, H2′),
5.61 (br, Hu). 19F NMR (CD2Cl2, 298 K, 376.65, J in Hz): δ
-75.10 (d, 1JFP ) 707.9, PF6-). 3BPh4: 1H NMR (CD2Cl2, 298
Experimental Section
RuCl3‚3H2O, ethylenediamine, N,N′-dimethylethylenedi-
amine, and N,N,N′,N′-tetramethylethylenediamine were pur-
chased from Sigma. [Ru(η6-arene)Cl2]2 was prepared according
to Benneth et al.35 Compounds 1-3 were prepared under
nitrogen using standard Schlenk techniques. Solvents were
freshly distilled (hexane with Na, Et2O with Na/benzophenone,
MeOH with CaH2, CH2Cl2 with P2O5) and degassed, by many
gas-pump-nitrogen cycles, before use.
3
K, 400.13 MHz, J values in Hz): δ 1.30 (d, JH7-H6 ) 6.94,
H7), 2.07 (s, H5), 2.18 (m, H8), 2.47 (m, H8), 2.77 (s, Med),
3.03 (sept, 3JH6-H7 ) 7.0, H6), 3.08 (s, Meu), 5.16 (d, 3JH3-H2
)
3
3
6.41, H3), 5.27 (d, JH2-H3 ) 6.41, H2), 6.93 (t, Jp-m ) 7.21,
p), 7.07 (t, 3Jm-o,p ) 7.34, m), 7.37 (br, o). 3PF6: 1H NMR (CD3-
3
OD, 298 K, 400.13 MHz, J in Hz): δ 1.31 (d, JH7-H6 ) 6.88,
H7), 2.25 (s, H5), 2.46 (m, H8), 2.57 (m, H8), 2.87 (s, Med),
3.08 (sept, JH6-H7 ) 6.79, H6), 3.38 (s, Meu), 5.49 (d, JH2-H3
3
3
One- and two-dimensional 1H, 13C, 19F, and 31P NMR spectra
were measured on Bruker DPX 200 and DRX 400 spectrom-
eters. Referencing is relative to TMS (1H and 13C), CCl3F(19F),
) 5.84, H2), 5.87 (d, JH3-H2 ) 5.71, H3). 19F NMR (CD3OD,
3
1
298 K, 376.65, J in Hz): δ -75.10 (d, JFP ) 707.9, PF6-).
T1 Measurements. The longitudinal relaxation times for
the 1H nuclei were measured by the standard inversion
recovery method on solutions of complexes 1PF6 in methylene
(35) Benneth, M. A.; Smith, A. K. J. Chem. Soc., Dalton Trans. 1974,
233.

3486 Organometallics, Vol. 24, No. 14, 2005
Zuccaccia and Macchioni
chloride-d2 at 296 K. A total of 15 experiments, having the
relaxation delay ranging from 0.001 to 15 s, were acquired,
each of them consisting of 32 scans. The total relaxation delay
between two consecutive scans was 15 s. All single experiments
were Fourier transformed. Frequency domain spectra were
processed with a standard T1/T2 software package available
on Bruker spectrometers to extract the relaxation parameters.
Important T1 values are 0.5 s for Hu, 0.6 s for Hd, 2.2 and 2.4
s for 2 and 3, respectively, 0.8 s for 8, 2.2 s for 6, 1.7 for 5-Me,
and 1.2 s for 7-iPr.
K)19,41 under the exact same conditions as the sample of
interest. From the measured Dt, rH was determined by
introducing into eq 3 the van der Waals radius of the solvent
as rsolv and the solution viscosity value corrected to take into
account both the actual temperature17 and the deuteration of
the solvents.18
(b) TMSS Solutions. Dt for TMSS and solvents that are
the sample (sa) and the internal standard (st), respectively,
were evaluated as described in (a). From eq 1 the ratio of the
Dt values for TMSS and solvent, equal to the ratio of the slopes
(m) of the straight lines coming from plotting log(I/Io) vs G2,
is
NOE Measurements. The 1H-NOESY36 NMR experiments
were acquired by the standard three-pulse sequence or by the
PFG version.37 Two-dimensional 19F,1H-HOESY NMR experi-
ments were acquired using the standard four-pulse sequence
or the modified version.38 The number of transients and the
number of data points were chosen according to the sample
concentration and to the desired final digital resolution.
Semiquantitative spectra were acquired using a 1 s relaxation
delay and 800 ms mixing times. Quantitative 1H-NOESY and
19F,1H-HOESY NMR experiments were carried out with a
relaxation delay of 10 s and a mixing time of 0.15 s (initial
rate approximation).39
Dsta cst rsHt
mst Dstt csa rsHa
msa
)
)
) f(rsolv,rsHa,rHst)
(6)
The latter circumvents the dependence of the Dt values on
temperature, solution viscosity, and gradient calibration. From
eq 6 the denominator csa rHsa can be determined since the msa/
mst ratio was measured and cst rsHt was evaluated as described
in (a). Since csa depends on rsolv and rHsa according to eq 2 and
rsolv was measured in point (a), rHsa was determined graphically
solving the equation (Supporting Information)
1
PGSE Measurements. H and 19F PGSE NMR measure-
ments were performed by using the standard stimulated echo
pulse sequence40 on a Bruker AVANCE DRX 400 spectrometer
equipped with a GREAT 1/10 gradient unit and a QNP probe
with a Z-gradient coil, at 296 K without spinning. The shape
of the gradients was rectangular, their duration (δ) was 4-5
ms, and their strength (G) was varied during the experiments.
All the spectra were acquired using 32K points and a spectral
width of 5000 (1H) and 18 000 (19F) Hz and processed with a
line broadening of 1.0 (1H) and 1.5 (19F) Hz. The semilogarith-
mic plots of ln(I/I0) vs G2 were fitted using a standard linear
regression algorithm; the R factor was always higher than
0.99. Different values of ∆, “nt” (number of transients), and
number of different gradient strengths (G) were used for
different samples. For example, for the measurements of 3PF6
0.7 mM in methylene chloride-d2, the “nt” values were 480 (1H)
and 960 (19F) with a total acquisition time of 18 (1H) and 24
(19F) h. 5, 7, Me, o-H, m-H, and fluorine resonances were
investigated as usual.
sa
6
csa rsHa
)
r
(7)
H
2.234
rsolv
1 + 0.695
sa
(
)
[
]
rH
(c) 1-3 Solutions Containing TMSS. Dt for 1-3 com-
pounds, TMSS, and solvents were evaluated as described in
(a). The correction factor for the 19F-PGSE measurements was
[γ(1H)/γ(19F)]2 ) 1.129. The rH values for compounds 1-3 were
estimated as described in (b), but in these cases TMSS and/or
the solvent itself served as internal standard.
The van der Waals volume of TMSS was computed from the
crystal structures42 using the software package WebLab View-
erLite 4.0. The volume of ion pairs was determined from the
crystal structure of 1PF6; for the other complexes, we added
the volume of the methyl group according to Bondi’s methodol-
ogy.6
The uncertainty in measurement was estimated by deter-
mining the standard deviation of m by performing experiments
with different ∆ values. The standard propagation of error
analysis gave a standard deviation of approximately 3-4% in
hydrodynamic radii and 10-15% in hydrodynamic volumes
and aggregation numbers N.
The dependence of the resonance intensity (I) on a constant
diffusion time and on a varied gradient strength (G) is
described by eq 5:
I
I0
δ
3
ln ) - (γδ)2Dt ∆ - G2
(5)
(
)
Acknowledgment. We thank the Ministero
dell’Istruzione, dell’Universita` e della Ricerca (MIUR,
Rome, Italy), Programma di Rilevante Interesse Nazio-
nale, Cofinanziamento 2004-2005 for support.
where I ) intensity of the observed spin-echo, I0 ) intensity
of the spin-echo without gradients, Dt ) diffusion coefficient,
∆ ) delay between the midpoints of the gradients, δ ) length
of the gradient pulse, and γ ) magnetogyric ratio.
(a) Pure Solvents. The diffusion coefficient Dt, which is
directly proportional to the slope of the regression line obtained
by plotting log(I/Io) vs G2 (eq 5), was estimated by measuring
the proportionality constant using a sample of HDO (0.04%)
in D2O (known diffusion coefficient in the range 274-318
Supporting Information Available: Data of PGSE NMR
measurements (Table S1-S27), dependence of N+ and N- on
the concentration for 3PF6 and 3BPh4 (Figures S1, S2),
dependence of crH on rH in methylene chloride-d2 (Figure S3),
representative two-dimensional experiments for the intramo-
lecular and interionic characterization of compounds 1-3
(Figures S4-S8). This material is available free of charge via
the Internet at http://pubs.acs.org.
(36) Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. J. Chem.
Phys. 1979, 71, 4546.
(37) Wagner, R.; Berger, S. J. Magn. Reson. A 1996, 123, 119.
(38) Lix, B.; So¨nnichsen, F. D.; Sykes, B. D. J. Magn. Reson. A 1996,
121, 83.
OM050145K
(41) Tyrrell, H. J. W.; Harris, K. R. Diffusion in Liquids; Butter-
worth: London, 1984.
(42) Dinnebier, R. E.; Dollase, W. A.; Helluy, X.; Ku¨mmerlen, J.;
Sebald, A.; Schmidt, M. U.; Pagola, S.; Stephens, P. W.; Van Smaalen,
S. Acta Crystallogr. 1999, B55, 1014.
(39) Neuhaus, D.; Williamson, M. The Nuclear Overhauser Effect
in Structural and Conformational Analysis, Wiley-VCH: Weinheim,
2000; Chapter 4.
(40) Tanner, J. J. Chem. Phys. 1970, 52, 2523.